Limits and Uses of Dynamical Predictions of Meteorological Drought

The overall technical capabilities now exist to make real time, seasonal drought forecasts on a near global scale, but how skillful are such predictions? In this talk the skill of seasonal drought indicator predictions based on a combination of real time observations and dynamical model seasonal forecasts is first evaluated over the US and Mexico. The relative contributions of predictive skill from sea surface temperatures and initialed land surface and atmospheric conditions is discussed relative to baseline predictability resulting from the inherent persistence of the indicators. Web-based tools which display such predictions are then briefly described. Finally, the challenges in using such predictions in decision-making settings is described. In many applications, more detailed or tailored information is desired. Examples of the latter are based on IRI-related projects on fire early warning in Kalimantan, food security outlooks in East Africa and research towards drought early warning in the agriculture sector in the Philippines and Sri Lanka.